DSXchange

PX does not have any native SAP 'plug-ins' that I know of. I do not have any experience in using PX with SAP.

Any DSX'ers with experience with SAP and PX?

- BP

Seems like your connection is timing out. What is the maximum time you can connect without activity to your mainframe?

- BP

Depends... Do you back up your PX server? You could possibly restore the RT_BP* files for the jobs off a backup.

If you don't back up the PX server or if you don't use version control with the PX GUI, you're going to have a hard time recovering deleted jobs.

- BP

Have you tried doing the same thing with a simple SQL script? This would isolate the issue to an Oracle error.

- BP

What is the problem you are having? Does PX not recognize the "" as a NULL?

The upgrade from 6x to 7x shouldn't be a problem. Just import the 6x dsx file into 7x. As for moving from loading data using sqlldr vs. using PX, it is as Ray said. Write the data out to a flat file and then call the sqlldr script outside of PX and give the sqlldr command the name of the flat file y...

You could add a record-type tag to each record and then just cat your files together on UNIX..

-BP

No more 'filter' option in 7.0? That's not so good. I can't think of any way to do it besides gunzipping the data outside of the PX job...

If you end up asking the Ascential folks what to do and they give you a good answer, please post the solution here..

Thanks,
BP

Rstone,

Did you resolve this issue? If so, what was the resolution?

Your descriptions of the Basic Transform in PX make it seem like the Basic Transform is buggy.. Care to comment?

I agree with Ray. If you are not able to solve the problem, could you post the generated basic code?

Thanks
BP

You can execute a user-defined sql statement within some of the PX database stages. You could try executing all three inserts in the user-defined sql section and put a commit at the end.

If you're pushing a large amount of data, Peter Nolan's suggestion is a good idea.

- BP

Or you can use the UNIX utility called sed. It is well documented on the web.

-BP

This should be a fairly straightforward constraint.

For outputA, set the constraint to COLA > COLB. For outputB, set the constraints to NOT(COLA > COLB).

Is this what you are doing?

-BP

.psds files are parallelized SAS datasets that have been "PX"d. You can't use them with native SAS, only with PX. If you do not want a parallel dataset, write the output of the stage to the Sequential File stage and create the dataset as a regular UNIX file.

-BP

Do OUTROWNUM and INROWNUM work in PX's transform? If so, that would be the way to get a row count. For a total record count of a dataset, use a 'dsrecords'.